Seasonal circulation of influenza A viruses (IAVs) annually causes substantial morbidity and mortality and has negative economic impacts estimated in the billions of dollars. Influenza pandemics caused by the emergence of novel, zoonotic IAVs pose an even greater threat. For decades, a key idea in influenza epidemiology has been that all humans would lack immunity against a novel influenza subtype that emerged from animals to cause a pandemic. However, our recent study showed that large parts of the human population actually have strong, pre-existing immunity against influenza A/H5N1 and A/H7N9, two avian subtypes of great concern for pandemic emergence in humans. Protection against particular subtypes with pandemic potential is predictably distributed across birth years, and occurs in individuals who were first exposed during childhood to an influenza A virus with hemagglutinin antigens (HA) in the same genetic group as the novel, emerging subtype of interest. The follow-up studies proposed here aim to (1) use reconstructed HA imprinting patterns to forecast age distributions of severe infection in future pandemics and (2) determine whether broadly-protective childhood HA imprinting affects the epidemiology of seasonal influenza subtypes A/H1N1 and A/H3N2. Aim 1 - pandemic influenza. We aim to determine whether, and by what factor, protective childhood HA imprinting reduces susceptibility to infection with novel IAVs, transmissibility of mild infections that do occur in protected individuals, or both. To estimate the size and significance of each possible effect, we will apply multitype branching processes and statistical inference to data on all known cases of A/H5N1 and A/H7N9. The results of this analysis will improve the accuracy of existing models, which can use information on birth year-specific childhood HA imprinting to forecast age distributions of severe infection in future pandemics. These forecasts can inform pandemic preparedness and response strategies, and can help target limited treatment resources toward high-risk age groups. Aim 2 ? seasonal influenza. Two IAV subtypes, H1N1 and H3N2, have circulated seasonally in humans since 1977, however H3N2 infections cause the vast majority of influenza-related fatalities in elderly adults. Mismatched childhood imprinting may contribute to increased H3N2 risk in the elderly, but observed differences in H3N2 vs. H1N1 risk could also be explained by higher intrinsic virulence of the H3N2 virus, or by other conventional, age-specific epidemiological factors. To determine whether, and how strongly, mismatched childhood imprinting contributes to increased H3N2 risk in elderly cohorts, I will analyze data on over 18,000 cases of seasonal influenza using a combination of statistical analyses, mathematical models and model comparison. This analysis will inform whether and how H3N2 vs. H1N1 risk in the elderly should be expected to change over time, as birth cohorts with different histories of childhood HA imprinting become older.